First-principles study of AlN nanosheets with chlorination

Abstract

Based on first-principles calculations, we study the effects of the chlorine atoms on electronic and magnetic properties of AlN nanosheets (NS). We find that both the bare and fully-chlorinated AlNNRs demonstrate semiconducting behavior, while the half-chlorination on surface Al sites leads to the semiconductor-ferromagnetism transition. More interestingly, the chlorination on surface Al sites in monolayer and bilayer AlNNSs demonstrates the half-metallic ferromagnetic (FM) behavior with 100% spin-polarized currents at the Fermi level, suitable for applications in spintronics at the nanoscale.

---

I. Introduction

Graphene, a two-dimensional (2D) honeycomb-like carbon network, has attracted much attention since it was synthesized in 2004.¹ It is noticeable that the electronic band structure of graphene has the linear dispersion character at the Fermi level (E_F) with gapless excitations, i.e., carriers in graphene behave like massless Dirac particles.² These unique properties revealed in graphene are expected to be used in the next-generation nanoelectronics in high-speed switching devices.^3–5 More recent works^6,7 reported that the foreign atom adsorption on 2D graphene can be achieved, leading to prospects for application in spintronics due to its room-temperature ferromagnetic (FM) order. Nevertheless, graphene suffers from several problems such as toxicity, difficulty in processing, and incompatibility with current silicon-based electronic technology.

The intense research dedicated to graphene generally triggered exploration into other 2D graphene-like networks including SiC, ZnO, and BNNSs,^8–10 which is considered of particular interest as a nanoelectronic device. The 2D BNNS has been exfoliated, which is a prerequisite for developing the full potential of h-BN in applications ranging from electronics to energy storage.¹¹ The problems related to geometric stability and electronic properties of 2D NSs of other III-Nitrides, including AlNNS, are recently becoming a “hot” research topic, which can expand the range of possible applications of III-Nitrides and open new perspectives for miniaturization in engineering functional nanodevices. Recently, the single crystalline AlNNSs has been successfully fabricated by a vapor-phase transport method, using Al powder and ammonia as the source materials.¹² Further examinations of the transmission electron microscopy (TEM) and scanning electron microscopy (SEM) demonstrated that the fabricated AlNNSs are uniform and smooth. Theoretically, Peng et al.¹³ reported the mechanical properties of AlNNS by density-functional theory (DFT) calculations, and found that the tunable sound velocities have promising applications in nanowaveguides and surface acoustic wave sensors. Rastegarn et al.¹⁴ found that the AlNNS can selectively detect NO₂ molecules in the presence of NH₃ molecules on AlNNS surface. Jiao et al.¹⁵ investigated the character of adsorption of CO₂ and nitrogen on single-layer AlN nanostructures, suggesting the potential application of AlNNS for CO₂ capture and storage.

Currently, there is an urgent interest in modifying 2D materials by the foreign atom adsorption to realize FM order in spintronics. For example, some previous results on graphene¹⁶ and silicene¹⁷ have shown the possibility of realizing intriguing magnetic orders. However, despite the progress in the fabrication of such AlNNS, less work on its magnetic properties has been reported on AlNNS. In our recent works,¹⁸ we investigated the electronic and magnetic properties of AlN nanostructure decorated with hydrogen atoms, and found the intriguing long-range FM orders. In this work, we further perform DFT simulations to investigate how to tune the magnetic properties of multi-layer (ML) AlNNSs by chlorine atom adsorption on AlNNS surface. One can see that the band gap of AlNNS can be significantly tuned when it is decorated with chlorine atoms. Of particular interest is the realization of the long-range FM order in such a 2D chlorinated AlNNS, which may open a new route to AlN nanostructures in spintronics.

II. Computational details

All DFT calculations are performed with spin-polarized plane-wave method implemented in the Vienna ab initio simulation package (VASP),^19,20 to investigate their geometric, electronic, and magnetic properties of the pristine and chlorinated ML AlNNSs. We use the projector augmented wave (PAW) potentials²¹ under the generalized gradient approximation (GGA) to describe the exchange and correlation interaction and a 450 eV cutoff energy for the plane-wave basis set is used. We employ the pseudo-potentials with 3s²3p¹, 2s²2p³, and 3s²3p⁵ valence electron configurations respectively for Al, N, and Cl atoms. Following the Monkhorst–Pack scheme,²² Brillouin zone integration is carried out at 9 × 9 × 1 special k-grids, and 15 × 15 × 1 k-points are used to obtain the electronic properties. The structural optimizations are performed using the conjugate gradient scheme until the maximum H–F force is smaller than 0.01 eV Å⁻¹.

Since the AlNNSs are cut from wurtzite bulk phase, the benchmark computations are performed for bulk model of wurtzite AlN. The optimized lattice parameters of bulk AlN are a = 3.087 Å, c = 4.962 Å. The calculated Al–N bond length is 1.872 Å, and Al–N–Al bond angles are 108.32°, all in good agreement with experimental values and other theoretical results.^23–28 The bulk AlN has an indirect band gap of 4.1 eV at the Γ point along Γ–Z direction in reciprocal spaces, close to the previous GGA results.²⁹ However, it is 0.8 eV higher than the value calculated with FLAPW method.³⁰ The possible reason can be attributed to the adopted different exchange and correlation potentials. It is well known that GGA typically underestimates the energy gap. However, since energy gaps are not the main focus of our work, the basic physics reported here should not be changed by using the GGA approach.

III. Results and discussion

We firstly construct 2D single-layer (SL) AlNNS normal to the (0001) surface of 3D bulk wurtzite structure, as shown in Fig. 1(a) and (b). The atomic structure of 2D AlNNS is similar to hexagonal structure of graphene,^1,2 excepted that the constituent atoms of the former are from III and V columns of the Periodic Table. After structural relaxation, it transforms from initial wurtzite configuration to a sp² hybridized planner graphitic-like structure, where the relaxed Al–N bond length of 1.798 Å is agreement with the previous theoretical value.³¹ Since the sp² bonding in 2D honeycomb structure is stronger than tetrahedrally coordinated sp³ bonding in 3D bulk, the decreases of Al–N bond length is not an exception, which is also observed in C, BN, and SiC honeycomb structure.³² Specifically, the Al–N bond length is 1.793 Å, which is ∼25% longer than the case of B–N bonding in BNNS. The N–Al–N and Al–N–Al angles are 120°and all Al and N atoms are within one plane. However, owing to the electronegativity difference between Al and N atoms, there are some electrons transferred from Al to N (0.62 e). As a result, in contrast to purely covalent bond in graphene, the bonding between Al and N gains an ionic character. In the cases of ML-AlNNSs [Fig. 1(c) and (d)], with the number of AlN layers (N) increasing, they always preserve the planner configuration if N < 5. Recently, many works reported the validity of pseudo-potentials on low-dimensional structures, such as hydrogenated graphene,³³ silicene,³⁴ BN,³⁵ and ZnO.³⁶ Also, in our recent work, we successfully predict the long-range FM order in 2D silicene.¹⁷ All these works suggest the availability of our computational approaches.

Fig. 1 (a) Top view and (b) side view of a SL-AlNNS. The rhombus plotted in dashed line represents unit cell of AlNNS. (c) and (d) present top view and side view of a four-layer AlNNS before optimization.

Examinations of band structure of the ML-AlNNSs indicate that they all exhibit semiconducting behaviors, with the band gap increasing monotonically as a function of the number layer N. In Fig. 2 we display the representative band structure and corresponding density of states (DOS) of SL-AlNNS with the band gap of 2.93 eV. One can see that the contribution of N-2p_z is pronounced for the filled band at the valence band maximum (VBM), while the conductor band minimum (CBM) comes from Al-3p_z states. The formed π- and π*- bands of 2D AlNNS which cross at K- and K*-points of the BZ in graphene open a gap in 2D AlNNS as a bonding and antibonding combination of N-2p_z and Al-3p_z orbitals. When further increases N of AlNNSs, the bands of each layer are not fully degenerated due to the layer–layer interaction, thus changes the band gap of AlNNS significantly.

Fig. 2 The calculated band structure and corresponding DOS SL-AlNNS. The horizontal dotted line in (a) and the vertical dotted line in (b) refer to the E_F, respectively.

Generally, the Al and N atoms in AlNNSs are more reactive than those of bulk AlN due to their dangling bonds on the surface atoms. To stabilize the plannar configuration, the conventional approach is decorating the surface Al and N atoms with the foreign atoms to saturate the dangling bonds, which can also be achieved by applying an external electric field perpendicularly to AlN (0001) surface. Different from the equivalent carbon atoms in graphene, AlNNS has two different kinds of atoms, and thus there are four different adsorbing configuration, i.e., top Al (T_Al) or N (T_N) sites, bridge site (B), and hollow site (H), as shown in Fig. 3(a). The relative stabilities of the adsorbing configurations is determined from the formation energy which is defined as E_f = E(AlN) + 1/2nμ_Cl − E(AlN@Cl_n), where E(AlN@Cl_n) and E(AlN) are the total energies of AlNNSs with and without the chlorine atoms, respectively. μ_Cl is the chemical potential of gases Cl₂, and n is the concentrations of Cl atoms in AlNNSs. Fig. 3(b) shows the calculated formation energy for four different sites. One can see that T_Al has the lowest formation energy, and thus the adsorption site T_Al would preferably be realized in experiments.

Fig. 3 The possible adsorbed sites and corresponding binding energies at different sites in AlNNS.

For the sake of comparison with T_Al, we first study the case of fully chlorination on both Al and N sties in AlNNSs, as shown in Fig. 4(a). The relaxed Al–N plane is found to be buckled with a height of 0.65 Å between Al and N atoms, attributed to transformation from bare AlNNS sp² hybridization to AlN@Cl_n sp³ hybridization between Al and N atoms. The Cl atoms are adsorbed on Al and N atoms with Cl–Al and Cl–N bond lengths of 2.07 Å and 1.79 Å, respectively. Obviously, the Cl–Al bond length is larger than that of Cl–N bonding because of the difference in bonding character. Fig. 4(b) presents the band structure of fully chlorinated AlNNS. It exhibits semiconducting character with a direct band gap of 0.52 eV, smaller than that of pristine AlNNS. This is because the adsorbed Cl atoms on both Al and N side lead to the strong hybridization at VBM mainly dominated by Cl-p, while the hybridization of Al-p and N-p orbitals is not affected in conduction band minimum (CBM).

Fig. 4 (a) Top and side view of geometry of fully-chlorinated AlNNS, and the corresponding band structure (b).

In the cases of half-chlorinated AlNNS, the structural relaxation of T_Al results in a trilayer configuration consisting of an Al plane sandwiched between Cl and N planes, as shown in Fig. 5(a). The distance d between Al and N planes is found to be 0.38 Å, while Cl–Al bond length d₀ is 1.71 Å. Also, we give the charge density difference (CDD) of T_Al on AlNNS in Fig. 5(b). The CDD is calculated by subtracting charge densities of free Cl atoms and AlN crystal from the charge density of 2D half-chlorinated AlNNS, i.e., Δρ = ρ(Cl@Al) − ρ(Cl) − ρ(AlN). One can see that high density contour plots around N atoms indicate no charge transfer from Al to N atoms significantly. The amount of transfer of charge is calculated by Bader charge analysis to be ΔQ = 0.01 electrons, which is consistent with CDD plots.

Fig. 5 (a) Top and side view of geometry of half-chlorinated AlNNS, and the corresponding charge–density difference ρ_Cl@AlN − ρ_Cl − ρ_Al − ρ_N of ML-Cl@AlN (b).

As discussed above, the chlorination on Al sites results in the strongly covalent bonds within Cl–Al atoms, which leads to sp³ hybridization in Al and N atoms, and thus no charge transfer occurs from Al to N, leaving electrons in N atoms unpaired. As a result, the N-2p states are spin-polarized with a net magnetic moment of 0.49μ_B, whereas the Cl atoms carry very small spins (0.08μ_B). To study the preferred magnetic coupling between N atoms, the energy difference, ΔE_FM = E_AFM − E_FM, between FM and antiferromagnetic (AFM) states, and spin polarization energy, ΔE_SP = E_NM − E_FM, between NM and FM states, are calculated. We find ΔE_FM = 0.18 eV and ΔE_SP = 0.84 eV, respectively. This indicates that T_Al prefers FM state. Noticeably, the N p states are rather extended. The extended p–p interaction between N atoms prefers a long-range FM coupling as found in graphene and BN sheets.³⁷

Fig. 6 presents the total and partial DOS of half-chlorinated AlNNS. It is obvious that the orbital hybridizations are mainly from N and Cl p orbitals. The large spin exchange interaction leads to spin-down p states move to E_F, while N 2p_z orbitals are pushed above E_F. Most importantly, the spin-down channel is metallic with the N p states crossing E_F, while the spin-up channel is semiconducting with a band gap of about 5.46 eV, exhibiting half-metallic properties, which is an ideal for high temperature operation up to room temperature.

Fig. 6 The calculated total DOS (a) and partial DOS (b) of half-chlorinated AlNNS.

In experiments, AlNNS prefers ML layers due to their low formation energies. Thus we further consider what would occur if increase the thickness N of AlNNS. The structural relaxations demonstrate that, similar to the pristine AlNNS, the half-chlorination on T_Al site leads to the reduction in the intralayer Al–N bonds, while the distance between the neighboring sheets becomes larger. Especially, the farther the Al–N layer is from Cl atoms, the larger the distance d is. As a result, these lead to the interlayer binding energies increases, and thus result in different magnetic properties from the case of AlN monolayer.

Finally, in Fig. 7 we present the band structure of ML-Cl@Al with N = 2 and 3, respectively. One can see that they are both spin-polarized clearly, suggesting magnetic properties. In the case of N = 2, the half-metallicity can be preserved and the half-metal gap is still large enough (0.52 eV) [Fig. 7(a)], while N = 3, the metallic behavior is observed since the bands in spin-up channel cross E_F, as shown in Fig. 7(b). The possible reason can be explained as follows; the Al atoms in inner layer provide electrons to transfer to N atoms at surface, reducing the number of unpaired N p electrons. Thus, the electrons of N in inner layer can't gain enough electrons from the neighboring Al atoms, leading to part of N p electrons in inner layer unpaired. In such a way, the local magnetic moments on N atoms are handed on layer by layer with increasing the number of Al–N layer. So, it is expected that the net magnetic moment in half-chlorinated AlNNSs is mainly constituted in unsaturated N p orbital in surface layer.

Fig. 7 The band structure in half-chlorinated AlNNSs for (a) N = 2 and (b) N = 3, respectively.

IV. Conclusion

We perform DFT calculations to investigate the geometric, electronic, and magnetic properties of 2D AlNNSs decorated with chlorine atoms. The formation energies analysis indicate that the chlorine atom prefer to adsorb on T_Al site, suggesting the adsorption on Al sites is the ground state. When the AlNNSs are fully chlorinated, they exhibit NM semiconducting behaviors with the band gap increasing significantly as a function of the number layer N. Half-chlorination on Al sites can result in a semiconductor-FM metal transition from SL to four-layer AlNNSs, respectively. More interestingly, we find that half-chlorinated SL and bilayer AlNNSs exhibit the half-metallic behavior, which is an ideal for practical applications. These findings in AlNNSs would pave a new way in designing spintronics devices in nanoelectronics.

Acknowledgements

This work was supported by National Natural Science Foundation of China (Grant nos 61076088, 11274143, 60471042 and 11304121), and Technological Development Program in Shandong Province Education Department (Grant no. J10LA16).

References

1. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666–669.
2. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, M. I. Katsnelson, I. V. Grigorieva, S. V. Dubonos and A. A. Firsov, Nature, 2005, 438, 197–200.
3. S. V. Morozov, K. S. Novoselov, M. I. Katsnelson, F. Schedin, D. C. Elias, J. A. Jaszczak and A. K. Geim, Phys. Rev. Lett., 2008, 100, 016602.
4. M. Otani, M. Koshino, Y. Takagi and S. Okada, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 81, 161403.
5. Y. Zhang, T.-T. Tang, C. Girit, M. C. Martin, A. Zettl, M. F. Crommie, Y. R. Shen and F. Wang, Nature, 2009, 459, 820–823.
6. D. C. Elias, R. R. Nair, T. M. G. Mohiuddin, S. V. Morozov, P. Blake, M. P. Halsall, A. C. Ferrari, D. W. Boukhvalov, M. L. Katsnelson and A. K. Geim, Science, 2009, 323, 610–613.
7. J. O. Sofo, A. S. Chaudhari and G. D. Barber, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 153401.
8. E. Bekaroglu, M. Topsakal, S. Cahangirov and S. Ciraci, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 81, 075433.
9. M. Topsakal, E. Aktürk and S. Ciraci, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 115442.
10. C. Tusche, H. L. Meyerheim and J. Kirschner, Phys. Rev. Lett., 2007, 99, 026102.
11. J. N. Coleman, et al., Science, 2011, 331, 568.
12. E. de Almeida, F. de Brito Mota, C. M. C. de Castilho, A. Kakanakova-Georgieva and G. Gueorguiev, Eur. Phys. J. B, 2012, 85, 48.
13. Q. Peng, X.-J. Chen, S. Liu and S. De, RSC Adv., 2013, 3, 7083–7092.
14. S. F. Rastegara, A. Ahmadi Peyghana, H. R. Ghenaatianb and N. L. Hadipoura, Appl. Surf. Sci., 2013, 274, 217–220.
15. Y. Jiao, A. Du, Z. Zhu, V. Rudolph and S. C. Smith, J. Phys. Chem. C, 2010, 114(17), 7846–7849.
16. J. Hong, E. Bekyarova, P. Liang, W. A. de Heer, R. C. Haddon and S. Khizroev, Sci. Rep., 2012, 2, 624.
17. C. W. Zhang and S. S. Yan, J. Phys. Chem. C, 2012, 116, 4163–4166.
18. C.-w. Zhang, P.-j. Wang, P. Li and F.-b. Zheng, J. Appl. Phys., 2012, 111, 043702.
19. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
20. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6, 15–50.
21. G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758–1775.
22. J. Monkhorst and J. D. Pack, Phys. Rev. B: Condens. Matter Mater. Phys., 1976, 13, 5188–5192.
23. A. Wright and J. Nelson, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 50, 2159.
24. A. Wright and J. Nelson, Phys. Rev. B: Condens. Matter Mater. Phys., 1995, 51, 7866.
25. O. Ambacher, J. Phys. D: Appl. Phys., 1998, 3, 12653–2710.
26. F. Bernardini, V. Fiorentini and D. Vanderbilt, Phys. Rev. B: Condens. Matter Mater. Phys., 1997, 56, R10024.
27. H. Schulz and K. H. Thiemann, Solid State Commun., 1997, 23, 815.
28. S. Strite and H. Morkoc, J. Vac. Sci. Technol., B: Microelectron. Nanometer Struct. – Process., Meas., Phenom., 1992, 10, 1237.
29. W. Jia, P. Han, M. Chi, S. Dang and B. Xu, J. Appl. Phys., 2007, 101, 113918.
30. V. I. Gavrilenko and R. Q. Wu, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 2632–2642.
31. H. Sahin, S. Cahangirov, M. Topsakal, E. Bekaroglu, E. Akturk, R. T. Senger and S. Ciraci, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80, 155453.
32. D. Golberg, Y. Bando, Y. Huang, T. Terao, M. Mitome, C. C. Tang and C. Y. Zhi, ACS Nano, 2010, 4, 2979–2993.
33. Y. Wang, Y. Ding, S. Shi and W. Tang, Appl. Phys. Lett., 2011, 98, 163104.
34. F. B. Zheng and C. W. Zhang, Nanoscale Res. Lett., 2012, 7, 422.
35. J. Duan, X. Huang and E. Wang, Mater. Lett., 2006, 60, 1918–1921.
36. Y. Wang, Y. Ding, J. Ni, S. Shi, C. Li and J. Shi, Appl. Phys. Lett., 2010, 96, 213117.
37. J. Zhou, Q. Wang, Q. Sun and P. Jena, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 81, 085442.

---